Instructions to use 0penAGI/0zhr with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use 0penAGI/0zhr with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf 0penAGI/0zhr
# Run inference directly in the terminal:
llama cli -hf 0penAGI/0zhr

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf 0penAGI/0zhr
# Run inference directly in the terminal:
llama cli -hf 0penAGI/0zhr

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf 0penAGI/0zhr
# Run inference directly in the terminal:
./llama-cli -hf 0penAGI/0zhr

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf 0penAGI/0zhr
# Run inference directly in the terminal:
./build/bin/llama-cli -hf 0penAGI/0zhr

Use Docker

docker model run hf.co/0penAGI/0zhr

LM Studio
Jan

vLLM

How to use 0penAGI/0zhr with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "0penAGI/0zhr"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "0penAGI/0zhr",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/0penAGI/0zhr

Ollama
How to use 0penAGI/0zhr with Ollama:
```
ollama run hf.co/0penAGI/0zhr
```

Unsloth Studio

How to use 0penAGI/0zhr with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for 0penAGI/0zhr to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for 0penAGI/0zhr to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for 0penAGI/0zhr to start chatting

How to use 0penAGI/0zhr with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf 0penAGI/0zhr

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "0penAGI/0zhr"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use 0penAGI/0zhr with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf 0penAGI/0zhr

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default 0penAGI/0zhr

Run Hermes

hermes

OpenClaw new

How to use 0penAGI/0zhr with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf 0penAGI/0zhr

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "0penAGI/0zhr" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Docker Model Runner
How to use 0penAGI/0zhr with Docker Model Runner:
```
docker model run hf.co/0penAGI/0zhr
```

Lemonade

How to use 0penAGI/0zhr with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull 0penAGI/0zhr

Run and chat with the model

lemonade run user.0zhr-{{QUANT_TAG}}

List all available models

lemonade list

Atomic Chat

0zhr

0zhr is an open conversational language model designed for reasoning, long-form dialogue, and tool-aware agent workflows.

Unlike lightweight chat models optimized only for response speed, 0zhr is designed to spend additional computation on planning, structured reasoning, and maintaining coherent conversations across longer interactions.

Source code & release: github.com/0penAGI/0pen — dataset pipeline, LoRA training script (lora.py), adapter weights, and GGUF conversion.

Highlights

🧠 Thinking / reasoning capable
🔧 Designed for tool calling workflows
💬 Strong conversational abilities
🌍 Multilingual (English + Russian)
📚 Long-form explanations
🤝 Natural dialogue style
⚡ Optimized for local inference

Philosophy

0zhr was trained with a focus on dialogue quality rather than benchmark optimization alone.

The goal is to produce responses that remain coherent during long conversations, reason before answering when appropriate, and integrate naturally into agent systems.

Capabilities

reasoning over multi-step problems
coding assistance
mathematics
writing
summarization
brainstorming
analysis
roleplay
tool-use planning
instruction following

Intended Use

0zhr is suitable for

local assistants
autonomous agents
research
coding
educational applications
creative writing
chatbot systems

Comparison

Model	Speed	Thinking	Tool Calls
0pen	⭐⭐⭐⭐⭐	❌	❌
0zhr	⭐⭐⭐⭐	✅	✅

0pen prioritizes latency.

0zhr prioritizes reasoning quality.

Training

0zhr is a LoRA fine-tune of Qwen3 4B (mlx-community/Qwen3-4B-4bit), trained on a dialogue-focused corpus emphasizing

coherent conversation
reasoning
reflective dialogue
instruction following
helpfulness
multilingual communication

The full training script (lora.py) is included in this repository. This release is the GGUF (Q4_K_M) export, fully compatible with llama.cpp and Ollama.

Example

User

Why does this mathematical proof fail?

0zhr

The proof becomes invalid when it divides by (a - b). Since the assumption is a = b, we have (a - b) = 0, making that step a division by zero. Every step after that no longer follows logically.

Run with Ollama

ollama create 0zhr -f Modelfile
ollama run 0zhr

Modelfile:

FROM ./0zhr_q4.gguf

SYSTEM """
You are 0zhr, created by 0penAGI. 
"""

PARAMETER temperature 0.55
PARAMETER top_p 0.9
PARAMETER num_ctx 32768

Run with llama.cpp

llama-cli -m 0zhr_q4.gguf -p "Hello, what can you do?" -n 256

Limitations

Like every language model,

can hallucinate
may generate incorrect factual information
should not replace professional advice
reasoning quality depends on prompt complexity

License

Apache-2.0

Made with ❤️ by 0penAGI

Downloads last month: 103

GGUF

Model size

4B params

Architecture

qwen3

Hardware compatibility

We're not able to determine the quantization variants.

View all variants

Model tree for 0penAGI/0zhr

Base model

Qwen/Qwen3-4B-Base

Finetuned

Qwen/Qwen3-4B

Adapter

(1102)

this model